A twisted-intramolecular-charge-transfer (TICT) based ratiometric fluorescent thermometer with a mega-Stokes shift and a positive temperature coefficient

Abstract

The fluorescence intensity of N,N-dimethyl-4-((2-methylquinolin-6-yl)ethynyl)aniline, exhibits an unusual intensification with increasing temperature, by activating more vibrational bands and leading to stronger TICT emissions upon heating in dimethyl sulfoxide. Based on the different temperature dependence at various wavelengths of the TICT fluorescence spectrum, this dye can be employed to ratiometrically detect temperature.

Full text

This journal is ©The Royal Society of Chemistry 201 4 Chem. Commun., 2014, 50, 15811--15814 | 15811
Cite this: Chem. Commun., 2014,
50,15811
A twisted-intramolecular-charge-transfer (TICT)
based ratiometric fluorescent thermometer with
a mega-Stokes shift and a positive temperature
coefficient†
Cheng Cao,
ab
Xiaogang Liu,*
bc
Qinglong Qiao,
b
Miao Zhao,
b
Wenting Yin,
b
Deqi Mao,
b
Hui Zhang*
a
and Zhaochao Xu*
b
The fluorescence intensity of N,N-dimethyl-4-((2-methylquinolin-
6-yl)ethynyl)aniline exhibits anunusual intensification with increasing
temperature, by activating more vibrational bands and leading to
stronger TICT emissions upon heating in dimethyl sulfoxide. Based
on the different temperature dependence at various wavelengths, as
shown in the TICT fluorescence spectrum, this dye can be employed
to ratiometrically detect temperature.
Fluorescent thermometers have been actively investigated for visual-
izing temperature distribution in micro-environments with high
spatial and temporal resolutions, i.e., in cells and micro-fluidic
devices.
1–8
Most of such thermometers are based on temperature-
sensitive organic dyes (such as rhodamine B) or inorganic com-
plexes of ruthenium or europium, whilst their emission intensities/
lifetimes decrease with rising temperature (owing to the thermal
activation of nonradiative de-excitation pathways), and this correla-
tion affords temperature information.
9–13
Recently, CdSe quantum
dots
14
and ZnO microcrystals
15
have also been reported as
nano-thermometers, which, however, still possess a negative
temperature coefficient. In contrast, fluorescent probes with a
positive temperature coefficient, whose emission intensities
increase with temperature, are particularly desirable, because
they can effectively suppress background interference at high
temperature; they can also work in conjunction with one fluoro-
phore with a negative temperature coefficient, where ratiometric
measurements of their emission intensities provide a built-in
correction and allow quantitative detection of temperature with
high sensitivity.
2,16–19
To this end, a number of polymer-based
fluorescent thermometers, which consist of a thermo-responsive
polymer and a polarity-sensitive fluorophore fragment (such as
benzofurazan), have been reported to have positive temperature
coefficients,
20–22
because polymers offer a decreasing micro-
environmental polarity around the fluorophore with a rise in
temperature. Unfortunately, these polymer-based thermometers
require complicated synthesis procedures; their temperature
sensing functionality only operates around the polymer phase
transition point (where significant micro-polarity changes occur)
and the corresponding working range is thus limited (i.e.,
o10 1C). A fluorophore with a positive temperature coefficient
has also been reported based on the thermally activated delayed
fluorescence of fullerene C
70
, whilst an increase in temperature
facilitates the back–forth transitions between the excited states
S
1
(singlet) and T
1
(triplet).
23
Nevertheless, such an effect is
vulnerable to oxygen quenching, posing a significant limitation
to its application environment. Consequently, the development
of new fluorescent thermometers, preferably with positive tempera-
ture coefficients, ratiometric measurements and a wide working
range becomes an important research target.
In this work, we propose and demonstrate a new design concept
to achieve a positive temperature coefficient in a fluorescent thermo-
meter, employing the twisted-intramolecular-charge-transfer (TICT)
emission of N,N-dimethyl-4-((2-methylquinolin-6-yl)ethynyl)aniline
(1; Scheme 1) in dimethyl sulfoxide (DMSO). TICT emission is
typically weak, because emission from the zero vibrational level of
the TICT state is forbidden; in contrast, the radiative transitions
from TICT to ground states are feasible via the assistance of higher
Scheme 1 The synthesis route of 1.
a
School of Pharmaceutical Engineering, Shenyang Pharmaceutical University,
Shenyang 110016, China. E-mail: zh19683@163.com
b
Key Laboratory of Separation Science for Analytical Chemistry,
Dalian Institute of Chemical Physics, Chinese Academy of Sciences,
Dalian 116023, China. E-mail: zcxu@dicp.ac.cn
c
Singapore-MIT Alliance for Research and Technology (SMART) Centre,
1 CREATE Way, Singapore 138602, Singapore. E-mail: xgliu83@gmail.com
†Electronic supplementary information (ESI) available: Synthesis, characteriza-
tion, experimental and computational details. See DOI: 10.1039/c4cc08010f
Received 10th October 2014,
Accepted 23rd October 2014
DOI: 10.1039/c4cc08010f
www.rsc.org/chemcomm
ChemComm
COMMUNICATION
Published on 29 October 2014. Downloaded by University of Cambridge on 19/11/2014 00:55:02.
View Article Online
View Journal
| View Issue

15812 |Chem. Commun., 2014, 50, 15811--15814 This journal is ©The Royal Society of Chemistry 2 014
and non-totally symmetrical vibrational bands of the TICT state.
24
These vibrational bands become activated upon heating. Con-
sequently, the TICT emission intensity will become intensified with
increasing temperature, provided that the increase in the TICT
emission rate exceeds that of non-emissive de-excitation.
Compound 1was synthesized in good yield, as shown in
Scheme 1 (ESI,†Fig. S1–S3). The UV-vis absorption solvato-
chromism of 1is very weak (Fig. 1a). For example, the peak
UV-vis absorption wavelength (l
abs
)of1changes from 332 nm in
n-hexane to 338 nm in DMSO, by only 6 nm. In contrast, the peak
emission wavelength (l
em
)of1in solution varies considerably from
non-polar to polar solvents (Fig. 1b). For instance, the peak l
em
is
registered at 380 nm in n-hexane and B550 nm in DMSO, with a
change of B170 nm. The substantial bathochromic shift in the
peak l
em
results in a mega-Stokes shift of B210 nm in DMSO.
A mega-Stokes shift is beneficial for reducing the overlap between
the UV-vis absorption and emission spectra of a fluorophore and
for minimizing the re-absorption of the emitted photons (via the so
called ‘‘inner-filter’’ effect), and thus improves the signal-to-noise
ratio of fluorescence imaging.
25
It is worth noting that the emission spectrum of 1in
n-hexane exhibits a clear shoulder; however, the fluorescence
spectra become featureless with large Stokes shifts, as solvent
polarity increases (Fig. 1b). Moreover, the transition energy of
the fluorescence (v
em
)andDf, the orientation polarisability
(eqn (1)), demonstrate an approximate linear relationship in
most solvents, except in n-hexane, as shown in a Lippert–Mataga
plot (correlation coefficient, r= 0.9287; Fig. 1c; Table S1, ESI†).
Interestingly, by re-plotting v
em
against Df0(eqn (2); which is
applicable when the TICT state is formed upon photoexcitation
of a dye), an even higher correlation coefficient of 0.9507 is
obtained (excluding data in n-hexane; Fig. S4, ESI†).
26
These
unusual spectral characteristics indicate the formation of the
TICT state in the excited state of 1.
Df¼
e1
2eþ1n21
2n2þ1(1)
Df0¼
e1
eþ2n21
2n2þ4(2)
where eand nrefer to the relative dielectric constant and the
refractive index of a solvent, respectively.
The formation of the TICT state in 1can be further ration-
alised via both viscosity-dependent emission measurements
and density functional theory (DFT) based calculations. In the
mixtures of ethanol and glycerol at various ratios, the UV-vis
absorption spectra of 1remain little changed (Fig. 2a). In
contrast, as more glycerol is added into the mixture, affording
a higher viscosity, the corresponding emission intensity of 1is
significantly enhanced, i.e., approximately three times from pure
ethanol to pure glycerol (Fig. 2b). Such viscosity-dependent fluores-
cence intensification is a typical characteristic of TICT emission.
24,27
Furthermore, DFT calculations show that the photoexcitation of 1
from S
0
to S
1
states mainly involves electron transitions from the
highest occupied molecular orbital (HOMO) to both the lowest
occupied molecular orbital (LUMO) and LUMO + 1 (Fig. 2c). Such
transitions result in substantial intramolecular charge transfer (ICT)
from the phenyl ring to the quinolone ring of 1. A large extent of ICT
is accompanied by significant molecular geometry relaxation and
drives the formation of the TICT state at the excited state of 1,
via the rotation of the dimethyl-amino group. The substantial
geometry relaxation in conjunction with the greatly increased
Fig. 1 (a) The UV-vis absorption and (b) fluorescence spectra of 1in
various solvents. (c) The Lippert–Mataga plot for the fluorescence transi-
tion energies of 1.[1]=10mM; the fluorescence spectra were obtained by
exciting the samples at their respective peak UV-vis absorption wave-
lengths. Note that the scales in (b) are arbitrary, owing to varied slit sizes in
use; please refer to Table S1 (ESI†) for exact fluorescence quantum yields in
various solvents.
Fig. 2 (a) The UV-vis absorption and (b) fluorescence spectra of 1in
ethanol–glycerol mixtures. [1]=5mM; the fluorescence spectra were
obtained by exciting the samples at 380 nm. (c) The frontier molecular
orbitals of 1in DMSO, obtained via DFT calculations.
Communication ChemComm
Published on 29 October 2014. Downloaded by University of Cambridge on 19/11/2014 00:55:02.
View Article Online

This journal is ©The Royal Society of Chemistry 201 4 Chem. Commun., 2014, 50, 15811--15814 | 15813
polarity of 1in the TICT state is responsible for the mega-Stokes
shift of 1and its large fluorescence solvatochromism.
25,28
The emission intensities of 1between B450 and B600 nm
exhibit an unusual positive temperature coefficient, with a
temperature coefficient of 0.5% per 1C, while that in the long-
wavelength region (4600 nm) decreases slightly with increasing
temperature in DMSO (Fig. 3a). Accordingly, the fluorescence
quantum yield of 1increases as temperature rises in DMSO
(Table S2, ESI†). This extraordinary effect can be rationalised
considering the nature of TICT emissions. As temperature
rises, more vibrational bands at higher energy levels in the
TICT state become active, which boosts the probability of
radiative electron transitions from TICT to ground states.
Not surprisingly, the associated peak emission wavelength of
1also displays a blue-shift from 551 nm at 25 1C to 538 nm at
65 1C, because TICT emissions associated with higher vibra-
tional bands contribute relatively more as temperature increases
(Fig. 3a).
24
It should be pointed out that the overall temperature depen-
dence of fluorescence intensity depends on two competing factors,
the rates of radiative TICT emission and non-emissive de-excitation.
Both rates increase with rising temperature. While the former is
relatively more significant, such as in DMSO, 1demonstrates a
positive temperature coefficient (Fig. 3a); in contrast, an inverse
trend is found in other solvents, such as ethyl acetate (EA; Fig. 3b),
owing to more substantial non-emissive de-excitation at high
temperature. Consequently, the emission quantum yield of 1 drops
with increasing temperature (Table S2, ESI†). In fact, the fluores-
cence intensities of most TICT dyes decrease at high temperature.
11
The different temperature dependence of TICT emissions at
varied wavelengths can be employed for ratiometric temperature
measurements (Fig. 3). The ratios of emission intensities at 500
and 600 nm in DMSO ([1]=5mM; l
ex
= 360 nm) afford an excellent
linearcorrelationwithtemperaturechanges,whichdemonstrates
the potential of using 1as a thermometer (Fig. 3d). A similar
temperature calibration curve has also been constructed in other
solvents, i.e., EA, although the overall emission intensity of 1
dropsastemperatureincreasesinthiscase(Fig.3e).Thereversi-
bility of this fluorescent system was also investigated by heating
and cooling it between 25 1Cand651Catastepchangeof101C
for 10 cycles in DMSO. The fluorescence intensities at different
temperatures remain consistent among all cycles, demonstrating
an excellent reversibility.
Compound 1can also work in conjunction with a tempera-
ture sensitive fluorophore, i.e., rhodamine 6G with a negative
temperature coefficient, to ratiometrically detect temperature
in DMSO (Fig. 3c and f). It is interesting to note that Fo
¨rster
resonance energy transfer (FRET) from 1to rhodamine 6G
occurs in this mixture, because the UV-vis absorption spectrum
of rhodamine 6G (represented by the grey shadowed area in
Fig. 3c) matches the emission spectrum of 1very well. As a
result, an emission peak at 567 nm, owing to rhodamine 6G,
became apparent even when the mixture was excited at 330 nm.
The TICT based thermometer of 1circumvents several limita-
tions of existing fluorescent temperature sensors. Fluorescence
lifetime-based thermometers require relatively long measurement
times and sophisticated equipment.
11
Fluorescence intensity-based
measurements are more straightforward. However, the emission
intensity also depends on the dye concentration and the illumina-
tion intensity. Careful initial calibration is thus unavoidable when a
single type of dye is employed in a thermometer. While a ratio-
metric thermometer using two types of dyes with distinct tempera-
ture coefficients provides a built-in correction, these two types of
dyes may photo-bleach at different rates, thus compromising the
reliability of the thermometer.
4
Moreover, the inhomogeneous
distribution of dyes (or concentration ratios) in a complex system
is a major concern for accurate interpretations of emission
intensity signals for both one-dye and two-dye thermometers.
In contrast, the thermometer based on 1offers a ratiometric
temperature sensing mechanism, while only one type of dye is
employed. This self-calibration effect makes the thermometer
insensitive to dye concentration variations and photo-bleaching
effects, thus greatly improving measurement accuracy. In addi-
tion, although the emission brightness of TICT dyes is normally
low, the quantum yields of 1are good, measured to be 55.2%
and 18.9% at 25 1C in EA and DMSO, respectively (Table S2,
ESI†). Since the temperature calibration curve of 1is solvent-
dependent and viscosity-dependent, this thermometer is most
suitable for a system with minimal solvent and viscosity varia-
tions. It can also serve as a calibration reference for existing one-
dye and two-dye thermometers to further gauge their accuracies,
i.e., in cells where the environment mostly consists of water.
Inconclusion,wehavedemonstrated a TICT induced emission
system with a positive temperature coefficient and a mega-Stokes
shift based on 1. This unusual positive temperature coefficient is
Fig. 3 Temperature dependence of the fluorescence spectra of (a) 1in DMSO
excited at 360 nm ([1]=5mM); (b) 1in EA excited at 360 nm ([1]=1mM); (c) the
mixture of 1and rhodamine 6G in DMSO excited at 330 and 540 nm,
respectively; [1]=5mM; [rhodamine 6G] = 1 mM. Temperature was varied from
25 to 65 1Catastepchangeof101C. The corresponding temperature
dependence of emission intensity ratios and the associated best-fit equations
of (d) 1inDMSOat500and600nm;(e)1in EA at 440 and 520 nm; (f) 1and
the rhodamine 6G mixture in DMSO at 500 and 563 nm.
ChemComm Communication
Published on 29 October 2014. Downloaded by University of Cambridge on 19/11/2014 00:55:02.
View Article Online

15814 |Chem. Commun., 2014, 50, 15811--15814 This journal is ©The Royal Society of Chemistry 20 14
realised by significantly activating vibrational bands of the TICT
state, which greatly facilitates TICT emissions with respect to non-
emissive de-excitation with a rise in temperature. Owing to
different temperature dependence of TICT fluorescence at various
wavelengths, the intensity ratios of these emissions can be used to
ratiometrically detect temperature. It is expected that this new
sensor design strategy will have a significant impact on the
developmentofratiometricfluorescent thermometers. However,
1haspoorsolubilityinaqueoussolutionandrelativelyshort
absorption wavelengths. Developing water-soluble and batho-
chromically shifted TICT dyes is the subject of our future work,
for measuring temperature in biological systems. The molecular
origins of the unusually high quantum yields and positive tem-
perature coefficients associated with the TICT emission of 1are
also currently under investigation.
Z.X. is grateful for the financial support from the National
Natural Science Foundation of China (21276251 and 21422606),
the Ministry of Human Resources and Social Security of PRC,
the 100 talents program funded by Chinese Academy of Sciences.
X.L. thanks the National Research Foundation of Singapore for
the SMART Scholar Postdoctoral Scholarship.
Notes and references
1 C. Gosse, C. Bergaud and P. Low, in Thermal Nanosystems and
Nanomaterials, ed. S. Volz, 2009, vol. 118, pp. 301–341.
2 E. J. McLaurin, L. R. Bradshaw and D. R. Gamelin, Chem. Mater.,
2013, 25, 1283–1292.
3 T. Ozawa, H. Yoshimura and S. B. Kim, Anal. Chem., 2013, 85,
590–609.
4 X. D. Wang, O. S. Wolfbeis and R. J. Meier, Chem. Soc. Rev., 2013, 42,
7834–7869.
5 Z. Yang, J. Cao, Y. He, J. H. Yang, T. Kim, X. Peng and J. S. Kim,
Chem. Soc. Rev., 2014, 43, 4563–4601.
6 K. Okabe, N. Inada, C. Gota, Y. Harada, T. Funatsu and S. Uchiyama,
Nat. Commun., 2012, 3, 705.
7 S. Kiyonaka, T. Kajimoto, R. Sakaguchi, D. Shinmi, M. Omatsu-Kanbe,
H. Matsuura, H. Imamura, T. Yoshizaki, I. Hamachi, T. Morii and
Y. Mori, Nat. Methods, 2013, 10, 1232–1238.
8 T. J. Johnson, D. Ross and L. E. Locascio, Anal. Chem., 2001, 74,
45–51.
9 N. Chandrasekharan and L. A. Kelly, J. Am. Chem. Soc., 2001, 123,
9898–9899.
10 S. Ryu, I. Yoo, S. Song, B. Yoon and J. M. Kim, J. Am. Chem. Soc.,
2009, 131, 3800–3801.
11 J. Feng, K. Tian, D. Hu, S. Wang, S. Li, Y. Zeng, Y. Li and G. Yang,
Angew. Chem., Int. Ed., 2011, 50, 8072–8076.
12 Y. Jiang, X. Yang, C. Ma, C. Wang, Y. Chen, F. Dong, B. Yang, K. Yu
and Q. Lin, ACS Appl. Mater. Interfaces, 2014, 6, 4650–4657.
13 V. F. Pais, J. M. Lassaletta, R. Fernandez, H. S. El-Sheshtawy, A. Ros
and U. Pischel, Chemistry, 2014, 20, 7638–7645.
14 L. M. Maestro, E. M. Rodriguez, F. S. Rodriguez, M. C. la Cruz,
A. Juarranz, R. Naccache, F. Vetrone, D. Jaque, J. A. Capobianco and
J. G. Sole, Nano Lett., 2010, 10, 5109–5115.
15 S. L. Shinde and K. K. Nanda, Angew. Chem., Int. Ed., 2013, 52,
11325–11328.
16 Y. Takei, S. Arai, A. Murata, M. Takabayashi, K. Oyama, S. Ishiwata,
S. Takeoka and M. Suzuki, ACS Nano, 2014, 8, 198–206.
17 T. Kan, H. Aoki, N. Binh-Khiem, K. Matsumoto and I. Shimoyama,
Sensors, 2013, 13, 4138–4145.
18 A. E. Albers, E. M. Chan, P. M. McBride, C. M. Ajo-Franklin, B. E.
Cohen and B. A. Helms, J. Am. Chem. Soc., 2012, 134, 9565–9568.
19 F. Ye, C. Wu, Y. Jin, Y. H. Chan, X. Zhang and D. T. Chiu, J. Am.
Chem. Soc., 2011, 133, 8146–8149.
20 C. Gota, K. Okabe, T. Funatsu, Y. Harada and S. Uchiyama, J. Am.
Chem. Soc., 2009, 131, 2766–2767.
21 S. Uchiyama, Y. Matsumura, A. P. de Silva and K. Iwai, Anal. Chem.,
2004, 76, 1793–1798.
22 S. Uchiyama, N. Kawai, A. P. de Silva and K. Iwai, J. Am. Chem. Soc.,
2004, 126, 3032–3033.
23 C. Baleiza
˜o, S. Nagl, S. M. Borisov, M. Scha
¨ferling, O. S. Wolfbeis
and M. N. Berberan-Santos, Chem. – Eur. J., 2007, 13, 3643–3651.
24 Z. R. Grabowski, K. Rotkiewicz and W. Rettig, Chem. Rev., 2003, 103,
3899–4032.
25 X. Liu, Z. Xu and J. M. Cole, J. Phys. Chem. C,2013,117, 16584–16595.
26 B. Valeur, Molecular Fluorescence: Principles and Applications, John
Wiley & Sons, 2013.
27 M. K. Kuimova, G. Yahioglu, J. A. Levitt and K. Suhling, J. Am. Chem.
Soc., 2008, 130, 6672–6673.
28 X. Liu, J. M. Cole and K. S. Low, J. Phys. Chem. C, 2013, 117,
14731–14740.
Communication ChemComm
Published on 29 October 2014. Downloaded by University of Cambridge on 19/11/2014 00:55:02.
View Article Online